Defensible or not defensible? Guard houses from Middle and Late Minoan Crete revisited, using GIS and Space Syntax

(1)

Defensible or not defensible?

Guard houses from Middle and Late Minoan Crete

revisited, using GIS and Space Syntax

(2)

(3)

Defensible or not defensible?

Guard houses from Middle and Late Minoan Crete revisited, using GIS and

Space Syntax

Sander L.G. Brabander

Master thesis in Classical and Mediterranean Archaeology Faculty of Archaeology, University of Leiden

Supervisor: Prof. Dr. J.L. Bintliff

Student number: 0601780 Course code: 1040X3053Y

(4)

(5)

3

Acknowledgements

It is a great pleasure to thank everyone who helped me write my thesis successfully. Tomas Alusik and his book “Defensive Architecture of Prehistoric Crete” for inspiring me to study the topic of Minoan guard houses. Quentin Letesson for a fruitful discussion on studying guard houses spatially. Prof. Apostolos Sarris for his help with GIS and my first DEM. Andrew Martindale for answering some questions related to my

implementation of his Defensibility Index. Hanna Stöger for her kind advice on applying Space Syntax analyses and for proofreading this thesis. Prof. John Bintliff for supervising me and, most of all, his patience. Finally, I would like to thank my parents, my sister and Claire van Driel for their everlasting support during the last year and a half.

(8)

6

Chronology Ceramic Period

high low sequence

3100 EM I Prepalatial Period

2700 EM IIA

2400 EM IIB (Court centred buildings?)

2200 EM III

2000 MM IA

1900 MM IB First Palace / Protopalatial Period 1800 MM II

1700 MM III Second Palace / Neopalatial Period 1700 1600 LM IA

1600 1500 LM IB

_{1400 LM II} Third Palace / Postpalatial Period 1400 LM IIIA1

LM IIIA2 1300 LM IIIB LM IIIC

1050 Sub-Minoan Sub-Minoan Period

Figure 1: Chronology of Minoan Crete. Dates are in years BC; high and low refer to the different

dating methods (see Shelmerdine 2008, 5-6). Based on Shelmerdine 2008, fig. 1.1, and Kyriakidis 2005, fig. 1.

(9)

7

1 Introduction

Until several years ago, I was fascinated by the idea of a peaceful Minoan society with its grand palaces, beautiful art and interesting lifestyles. However, the more I studied this great Cretan society, the more it became clear that this simple and perhaps naïve idea was not true at all. On the contrary, the strong and thick walls of barriers and buildings, and the collection of swords and daggers, among other things, have been interpreted as evidence of social unrest or even warfare. This was done under the assumptions, that strong walls must have a defensive function and swords are a symbol of the warrior and thus warfare. These assumptions will be discussed later in this thesis.

Some of the most interesting structures that have often been assigned a defensive function, are the Minoan guard houses. These rather small buildings were constructed and used between the early Protopalatial period and the Dark Ages. Their strong walls and their locations in the landscape are the main reasons why researchers have identified this type of structure as defensive and/or defensible. In this thesis, the main focus will be on the defensibility of these guard houses during the Minoan periods (see Figure 1 for a chronology of Minoan Crete). However, before discussing the specific research questions and how the research was conducted, it is important to give the necessary context, in which this research can be placed.

Arthur Evans was the first archaeologist to define Minoan society. His interpretations of how society on Crete functioned during the Bronze Age strongly influenced Minoan archaeology for at least until the second half of the twentieth century. Evans’ most important publication ‘The Palace of Minos’ (1921, 1928, 1930, 1935 and 1936) promoted his ideas of a peaceful society. Interestingly, in the years before the discovery of Knossos, Evans actually already published the finding of multiple strong walls, buildings he called fortresses, bastions, etc. (see Brown 1993). In more recent works these structures have been interpreted as signs of unrest and conflict rather than peace. Although Evans had recognized those architectural remains and even published a mosaic showing a fortified town, he held on to his idea of the Pax Minoica. This was also because Evans thought many of the fortifications around Knossos had been abandoned in the later Middle Minoan period, except for some guard houses watching over important roads (Evans 1928, 372).

During the 1930s Pendlebury made major contributions to Minoan archaeology and laid the basis for future research. Not only did he discover several important defensive sites on the island, such as Oreino Ellinika and Stavrochori Kastro (Pendlebury 1939, 385) he was also the first to properly describe the settlement patterns on Crete

(10)

8 during both the Minoan and the later Mycenaean periods (Nowicki 2000, 13). In a way, all later studies of settlement patterns on Crete can be seen as a continuation of his work. However, even though he had discovered more fortifications, the Minoans were still seen as a peaceful people.

During the 1950s and 1960s, large research projects were carried out by N. Platon, P. Faure and a British team including S. Hood, P. Warren and G. Cadogan. By this time, more and more evidence was found pointing to an overall less peaceful society. However, it was not until Alexiou (1979) openly doubted the peacefulness of the Minoan society, that a shift in Minoan archaeology started, gradually moving from the firm belief in a long period of peace and stability to ideas about social unrest, conflict and even warfare.

From about the 1960s onwards, warfare became a popular topic in anthropology. There it was suggested that warfare could be one of the reasons of how early states formed. Some scholars even consider warfare as one of the most important factors for the appearance of complex societies (Driessen 1999, 11-2). As with many other theoretical approaches, this anthropological theory was accepted by archaeologists as well, mainly because conflict at different levels can be a useful tool for explaining all kinds of events (Driessen 1999, 12). The combined efforts of archaeologists, anthropologists and historians have led to a better understanding of warfare in prehistoric and historic times (Parkinson and Duffy 2007, 116). From the late 1980s onwards the importance of warfare in Minoan archaeology steadily grew, with a peak in the late 1990s. This can be seen from the amount of research projects (e.g. the Minoan Roads Project, the project by Nowicki and the project by Schlager), publications (e.g. Nowicki 2000; Schlager 1997), and conferences relating to defensive architecture and warfare (e.g. POLEMOS).

Today it seems the interest in warfare and defensive/defensible architecture from the Minoan periods has died down, with only N. Schlager and his colleagues still actively studying the fortifications in mainly the Ziros region on Crete. However, in my opinion there are many questions still unanswered and further research is needed.

The above is a brief overview of the development within Minoan archaeology from the belief in peace and stability to a recognition of instability and possibly warfare. However, what is the current opinion held by Minoan scholars? What have we learned during this period of (active) discussion about conflict and social unrest in Minoan society?

For the Early Minoan periods, there is hardly any evidence available relating to conflict. However, there seems to have been more interest in settling on higher locations that were easier to defend and also harder to access (Nowicki 2000, 38). This has been

(11)

9 interpreted as evidence for increasing insecurity from the end of the Neolithic onwards (Alexiou 1979, 55).

Nowicki (2000, 38) recognizes at least three periods of conflict and/or crisis. The first being during the transition from the Final Neolithic to Early Minoan I, which was discussed briefly in the last paragraph. The second was during EM II and after the destructions of EM II, where it seems people again moved to more defensible locations. The third period of crisis, according to Nowicki, was during LM IB. During LM IB there was a series of events that more or less caused the destruction of the Neopalatial state.

Interestingly, although Nowicki identifies at least three periods of conflict, he left out the Middle Minoan periods from these three, while it was during these periods that many of the defensive architecture was constructed. Social inequalities started to appear on Crete near the end of the EM period and became more apparent from the MM period onwards. Apart from the development of a more hierarchical society, it is also possible that population movements, especially from the southern parts of Crete to the east, caused some form of conflict (Tsipopoulou 1999, 180). During MM IB/II there was an increase in sites that were fortified and/or located on defensible locations. Not only that, it is even suggested that some of them were part of a form of organized system of defences and fortifications (Nowicki 2000, 38). Especially between MM II and MM III there seems to have been a period of unrest. Defensible sites from MM II and later are distributed over several parts of the island. This may be evidence for specific zones where conflict or social unrest was present. Nowicki suggests these areas may be showing the final phase of the division of Crete into several political zones, a process which had started already during the EM II period (Nowicki 1999, 193).

During MM IIB the most important palatial centres of the Protopalatial period (Knossos, Malia and Phaistos), and many other settlements as well, were destroyed by a fire at almost the same time (Driessen and Macdonald 1997, 12). Driessen and Macdonald suggest the possibility of warfare of some sort, based on the fact that some settlements were immediately rebuilt after the destruction while others were left abandoned. It has also been suggested that earthquakes were the cause of this destruction, for example at Phaistos (Militello 2011, 239), but not everyone agrees with this (Bintliff 2012, 137). Some areas may actually have benefited from these destructions (Nowicki 2000, 32-3). This may be another reason to suggest conflict or social unrest. Both conflict and natural disasters may be possible explanations, or perhaps a combination of the two, although it could also have been large forest fires caused by drought and high temperatures.

After the destructions of MM IIB, it seems a new political and administrative organization emerged on Crete during MM III (Nowicki 2000, 38), possibly with

(12)

10 Knossos as controlling palatial centre (Driessen and Macdonald 1997, 12). The Neopalatial period seems to have been an era of relative peace. This is suggested by the abandonment of most of the defensible sites from earlier periods (Nowicki 2000, 32-3). This may indicate a nonexistence of unrest or conflict.

At the end of the Neopalatial period there were two major events that eventually caused the end of the Minoan palatial society. At the end of LM IA many sites were damaged by earthquakes and subsequently abandoned, but much was rebuilt soon after (Driessen and MacGillivray 2011, 267). The destructions of this period are usually ascribed to the earthquakes that preceded the eruption of the volcano of Thera and the eruption itself (Nowicki 2000, 34). Soles (1990, 322) has reported a layer of ash between the LM IA and LM IB levels of occupation. It is assumed that this layer of ash may have covered a large area of central and eastern Crete. This layer of ash has also been identified at Palaikastro, where the layers of pure volcanic ash could reach a thickness of up to 12 cm (Bruins et al. 2008, 202-3). It was long thought this may have caused agricultural problems, especially where the layer of ash was thick, but this is not necessarily the case. Depending on the season, the weather, the type of crops, amount of maintenance, etc. the effect of ash on soil differs (see also USGS 2009) Actually, a recent study suggests the sulphur and minerals in the ash that is found on Crete may have had a beneficial effect on plants (Pearson et al. 2009, 1212). Since a certain amount of sulphur is actually beneficial to the growth of wheat, there may not have been any agricultural problems at all.

Driessen and Macdonald (1997, 96) think the events surrounding the Thera eruption caused a collapse of the existing central authority (see also Driessen 2002, 251). They argue that the break-down of the system can be explained from the way people responded to the problems they were facing. They believe that in societies like the Minoan society, with a form of central authority, but where the (extended) family still has an important rol, people will first make sure their family members are safe when natural disasters happen. Based on the extended families, some form of regionalism exists in such societies. In crisis situations, the central authority loses its power, because groups of people will manage the situation by themselves, without a need for central authority. The only way for central authority to return to power is if it manages to reclaim control (Driessen and Macdonald 1997, 96).

The above is a rough sketch of what is assumed to have been the situation during the early LM IB period. This regionalisation seems to be supported by the appearance of Linear A tablets at many different locations, compared to the few locations they were found at before (Driessen and Macdonald 1997, 96; Hamilakis 2002, 194). This decentralisation does not, however, mean a decline. After the destructions related to the

(13)

11 Santorini eruption, much was rebuilt (Driessen and MacGillivray 2011, 267). LM IB was the height of the Minoan civilization. The marine style pottery that appeared in LM IB can be seen as evidence of this, although this style differed from earlier styles (Younger and Rehak 2008, 153). It was not until the end of LM IB that things changed drastically. However, evidence suggests this period was not as safe as LM IA. Buildings were modified to regulate and direct movement, and defences were added to settlements. It also seems food production and industry was now done inside settlements instead of in the hinterland. Old buildings that were destroyed at the end of LM IA were sometimes repaired and/or adapted to new uses, but the quality of the architecture was not as good as before. The lack of a central authority also made the hinterland more insecure, making communication within the island more difficult. Another strong clue for a sense of insecurity among the people, is the occurrence of a large number of bronze hoards at least during LM IB, but maybe earlier as well (Driessen and Macdonald 1997, 82-3).

The above gives some examples of evidence suggesting a period in which central authority was mostly lacking and people were feeling insecure. Defences around settlements, combined with the move of important assets like industry and food production to inside the perimeter of settlements, is a clear indication that the hinterland had become less secure. Industry, like metal working, was usually carried out outside of settlements, to prevent fires. The need must have been very high for the people to accept that risk over losing production facilities.

During LM IB there was another destruction phase. The evidence suggests a “violent destruction” that seems to have been aimed especially at important political and administrative centres (Nowicki 2000, 35). This selection indicates choice, which would point towards human agents at work causing these LM IB destructions. This has also been recognised by other scholars, who have noticed a kind of preferential treatment, where, within a site, only the elite houses were burned down, while other structures remained unharmed (Driessen 2002, 251). This is something that nature just cannot accomplish. Driessen (2002, 251-2) sees five reasons why there were humans causing havoc:

1. Preferential treatment, which has already been mentioned above. 2. Burning.

3. Plunder and malicious destruction.

4. Lack of reoccupation after the destruction.

5. Important facilities for water and storage were protected by defensive walls that restricted access to these facilities before the destructions took place. This suggests a threat that can be kept out...other people. Also, there are cases where

(14)

12 valuables had been hidden before the destruction. Again, this seems to suggest that the inhabitants were aware of an incoming threat.

If the destruction was indeed caused by other people, there are two options. Either there is internal conflict among the Minoan population, caused by social, economical or political problems, or there is an invasion by people from outside of Crete. Driessen and Macdonald prefer a combination of these two options (Driessen and Macdonald 1997, 109). Now there seems to be general agreement about an invading force attacking Crete. Still, some interpret the destruction of LM IB as the result of natural disaster, based on the amount of damage that was done. Nowicki does not agree with their standpoint, that this amount of damage could not have been done by a human agency. On the contrary, he argue the destructive power of man is similar if not worse than nature in some cases. He, too, agrees that most of the evidence seems to suggest that invaders, possibly from the mainland, were the agent responsible for the destructions (Nowicki 2000, 35). Although there is evidence suggesting the invaders came from the mainland, this is still open to debate.

Whatever the case may be, what is certain is that the events that occurred during LM IB had a dramatic impact on Minoan society. Not only was there a large decrease in population, which Nowicki theorizes may have been caused by invaders or raiders taking many prisoners (Nowicki 2000, 259-60), the Minoan Neopalatial society was also either destroyed or severely damaged.

However, Minoan society did not completely collapse. During LM II, which was still a relatively unstable period, Minoan society slowly recovered. Because of the influences from mainland Greece, this is often seen as the start of the Mycenaean period on Crete (Preston 2008, 311), but there is clear evidence for continuity as well (Driessen and Macdonald 1997, 110-1). During LM II-IIIB, the larger and more important towns, for example Palaikastro and Knossos, were reoccupied, but many other sites had been left abandoned after the destructions of LM IB (Nowicki 2000, 35). Based on the use and spread of LM II-IIIA pottery from Knossos, Driessen and Schoep (1999, 389) believe that it is possible that Knossos was in control of a much larger area than before. This is supported by Linear B tablets that list many names of places to the west of Chania and in the eastern part of Crete. They believe Knossos managed to gain control of such a large territory by using military power and the creation of a good communication network to lay a legitimate claim on the other areas of Crete (Driessen and Schoep 1999, 389; Preston 2008, 311, 6; Schoep 2007, 215). At the start of LM IIIA2 there seems to have been more destruction at Knossos and possibly at other sites as well, as suggested by evidence from Palaikastro. It is not known whether Knossos was (partly) restored afterwards or not (Nowicki 2000, 36).

(15)

13 The brief outline given above discussed the major crisis events throughout the Minoan periods. Hopefully it has become clear that the Cretan Bronze Age was not just peace and happiness, but that there were major periods of distress. Some of which were caused by natural disasters, others by people, either from the island itself or from the outside. In the next section I will discuss more in debt the fortifications that have been found on Crete, moving closer towards the topic of this thesis. Before continuing, however, there is the issue of warfare that needs a little more attention.

Conflict and social unrest are one thing, calling something warfare is a different matter. The term warfare has a certain baggage of modern assumptions going with it. Our knowledge of Aegean warfare is completely based on inference from different materials and some limited textual and iconographic evidence. Therefore it is more indirect than direct evidence. The way we look at warfare is also often affected by our own assumptions about warfare, that may lead to conclusions that may not be true at all (Krzyszkowska 1999, 489).

If we look at Minoan art, for example, there is one thing clearly missing, especially when comparing with art from other parts of the Mediterranean at the time. There is almost a lack of violence and scenes of warfare. This can either be explained by concluding that the Minoans did not have warfare, or if they did have warfare, they did not like to represent it in art (Gates 1999, 277). Because of all the other evidence, as discussed above, and the doubt that any society can have no conflict at all for multiple centuries, the second explanation is nowadays accepted. However, there is still some doubt if real warfare really existed, since there is almost no real evidence for warfare. Gates (1999, 277) gives five reasons for this doubt:

1. During the Neopalatial period, there are almost no fortified settlements.

2. Although weapons, mostly swords, have been found, they are still not common. 3. The lack of war scenes in art.

4. No written reports from that time period about warfare. 5. No burials with grave goods that can be related to warfare.

However, some of these can be explained relatively easy, while some cannot. For instance, the lack of fortified settlements in a large part of Crete during most of the Neopalatial period is still not understood. More is known about Minoan writing, however. Although the Minoans did have multiple writing systems, they were mostly used for record keeping (as far as we know), so the lack of written reports about warfare from the Minoans themselves does not per se mean anything special, because they did not write about others things either. There are no warrior-related symbols, such as weapons and helmets, on Linear A tablets, in contrast to Linear B tablets, where they are represented (Bintliff 2012, 149). Other complex Near Eastern cultures may have had no interest in

(16)

14 writing about warfare on Crete or they did not know. Either way, there are multiple reasons which could explain the lack of written reports. Interestingly, for all other East Mediterranean cultures where we have textual evidence of, warfare was very important.

A type of object that is often related to warfare is the sword. Of all the Bronze Age swords from the Aegean that have been catalogued, at least half are said to have been crafted on Crete. Together with the idea that many Mycenaean swords have actually been crafted in Minoan workshop, this seems to show a strong connection between Minoans and weapons (Peatfield 1999, 68). The high percentage of swords crafted on Crete is very interesting and seems to imply that weapons were very important for the Minoans in some way. Why else would they craft so many? However, although there is clear relation between swords and warriors in Aegean art in general, there are enough examples of swords of different types that clearly have been used for ceremonial purposes. Good examples of a ceremonial purpose are the swords found at Malia (Peatfield 1999, 69). Therefore, we must be cautious not the judge too soon when looking at the weapons that have been found, at least for the Early and Middle Bronze Age. For the Late Minoan periods, there is more evidence for the actual use of swords by warriors in art and graves (Peatfield 1999, 70; 2). A good example of this, is a rich grave that was discovered at Poros, the harbour town of Knossos, which not only included jewellery, but weapons as well (Driessen and Langohr 2007, 186).

From what has been argued above, we can say that conflict and crisis existed, but whether there was real warfare going on in Crete cannot be said with certainty. The existence of weapons is not evidence in itself for real warfare, and the lack of representation of warriors in most EM and MM art does not help us much either. During the LM periods, there is an increase in representations of warriors, and during LM II-IIIA the weapons and armour found in burials, combined with written records of weapon production, chariots, etc. on tablets, make it clear that warfare had become very important on Crete for elites to gain power (Driessen and Schoep 1999, 392-3). However, especially during LM II-IIIA these must have been Mycenaean influences.

Whatever the case may be, people felt the need to defend themselves during multiple periods of Minoan prehistory. This was either done by building a settlement on higher locations that may have been harder to access, by building fortifications of different kinds, or a combination of both. Guard houses, the main object of this thesis, are a type of fortification. However, they were part of a variety of fortifications and it is important to see them in relation to each other.

Fortifications can be divided into three categories: forts and guard houses, towers or bastions, and perimeter defence or enclosure walls (Zielinski 1998, 61). Zielinski and Schlager, among others, use the term ‘cyclopean’ for the strong and thick walls that are

(17)

15 used for the construction of most fortifications. However, based on this type of wall more categories can be defined. Schlager (2006, 371) defined six categories:”fortifications of settlements and towns”, “guard houses along Minoan roads”, “sanctuaries”, “town houses”, “central edifices in village-like settlements” and “farmsteads”. These six categories show that strong “cyclopean” walls were not only used for fortifications, but for other types of structures as well. This is important to keep in mind, because it means strong walls are not per se evidence of fortifications.

From the Neolithic through the Bronze Age, fortifications have been used to define and protect territories (Alusik 2007, 175). The amount of defensive architecture differed depending on the period. Especially in periods of crisis there seems to have been an increase in fortifications of different sorts. Two of the oldest examples of fortified sites are the bastion and acropolis at Livari, both dating to the Final Neolithic period. In both cases there was a perimeter wall surrounding the site (Schlager 2011, 272-3). Although there are some fortifications known from the early Prepalatial period, most seem to have been constructed from the late Prepalatial period onwards.

During EM III-MM I simple walls, consisting of large blocks of stones with smaller stones used as a filler, were constructed around hilltop villages for protection. Apparently, these were often placed strategically, for example at the most exposed part of the site, to control access from specific directions (Hayden 1988, 1-2). Nowicki (2000, 32) agrees that during this period there were more defensible sites than before. However, he asserts that real fortified sites are usually seen as a feature of MM II and (maybe) early MM III (Nowicki 2000, 32). This seems to fit with all the evidence that indicates an increase in defensive architecture starting from the early Protopalatial period onwards. The stronger concern with defence seems to coincide with a period of great social unrest and change, in which multiple palatial centres were developing and claiming territories of their own. Schlager proposes that these political powers wanted to define themselves and their territories, and that they saw the construction of ‘cyclopean’ buildings as a good method to make a clear statement of their power and status. Moreover, it may even have been the case that the opposing groups started to build defensive architecture as well (Schlager 2006, 376). Notwithstanding all this, there is still a lot of uncertainty about the function of many ‘cyclopean’ structures. For example, a large number of walls do not seem to have many defensive qualities at all. They functioned well enough as simple barriers and to keep animals and perhaps products safe, but had no real defensive value (Driessen and Macdonald 1997, 47).

Nevertheless, perimeter walls seem to have been the most common fortifications. Many were constructed around important sites or used as field boundaries. During the Protopalatial period, several large centres gained defensive walls around them, for

(18)

16 example at Petras and Aghia Photia. This is seen as evidence for conflict in the area (Tsipopoulou 1999, 185). On different sides around Malia, wall segments have been discovered as well, that have been interpreted as fortifications around this important town (Zielinski 1998, 63-4). However, Müller did already in the early 1990s doubt this, for three reasons: 1) the appearance and width of the segments differs greatly; 1) they do not form continuous lines; 3) in at least one case there are perperndicular walls attached to it (Müller 1992, 745). Palaikastro had fortifications as well, and the area south of Palaikastro, the territory of Kato Zakros, had a considerable number of fortifications in the countryside (Zielinski 1998, 452).

‘Cyclopean’ walls were either used on their own as periboloi, enclosure walls, or as part of another structure. This could be any type of structure, ranging from villas and farmsteads to shrines, towers and guard houses. Together these buildings and periboloi formed a complex defensive system (Schlager 2006, 370-1).

Of course walls were not the only type of fortifications. As was mentioned above, if only the structures are counted that really seem defensive, there were also towers, bastions and guard houses or forts. A good example of a combination of strong walls with a sort of bastions can be found at Aspro Nero. There, Schlager and his team have discovered at least two bastions connected to each other by walls. These could be dated, based on pottery finds, to the MM period and the site was probably still in use during the LM period (Schlager et al. in press). Guard houses will be discussed in more detail in a following section, since they represent an important type of ‘cyclopean’ buildings.

At the end of MM IIB, the wide destructions not only had an impact on the palatial centres and the settlements around them, but also on the use of defensive architecture, like guard houses, and many defensive sites were either abandoned or destroyed (Alusik 2007, 151). During MM III-LM I there is also evidence that defensive sites were being used, but most of them were founded earlier and were just being reused. It also seems that more settlements were now built on lower areas, suggesting a safer environment (Nowicki 2000, 33). During most of the Neopalatial period, there was a decrease in the use of fortifications around towns. A possible reason for this seems to be the network of guard houses that was spread over the countryside, especially near Zakros. These guard houses were used for watching the area for possible threats, making town walls no longer a necessity (Zielinski 1998, 524), but this is not certain at all. The change from city walls and guard houses protecting the main towns to a network of guard houses protecting the borders of territories, if that is indeed the case, may be a response to political changes. It seems plausible that when the internal conflicts were resolved and threats were almost only external, watching the borders became more important than controlling the interior (Alusik 2007, 134). This also explains why a town like Gournia,

(19)

17 which was founded during the Neopalatial period, does not have any fortification walls around it (Zielinski 1998, 528).

In LM I there was apparently a new need for fortifications. Important settlements like Achladia, Aghia Triada, Gournia, Petras, Palaikastro and Kato Zakros received enclosure walls. In several cases, for example at Gournia, one or more towers were added as well (Driessen and Macdonald 1997, 46-7). This suggests again a threat that needs to be kept out. Although the first major destruction phase in LM I was caused by the earthquakes related to the eruption of the Thera volcano and the eruption itself, it is also possible a human threat arrived in LM IB, as has been discussed earlier.

Next to the addition of walls and towers to these palatial sites, there is also evidence at several sites, for example at Knossos, for a change in accessibility. Changes to entrances caused people’s movement to be more controlled. People were actually forced to move a specific way to reach the interior. Driessen and Macdonald suggest this may either point to more social differentiation or to a need for controlling accessibility for safety reasons (Driessen and Macdonald 1997, 45-6).

After the destructions of LM IB, there was again a decline in the use of defensible sites and fortifications. One possible explanation is that the remaining inhabitants on Crete were pacified by invaders, removing the need for defences and defensible sites (Nowicki 2000, 36). However, since it is still not completely safe to assume there really was an invasion by raiders, there could be other reasons as well. Even if there were no invaders, or if they merely caused chaos and disorder and then left, the Minoan society may have been united in some way, which would also make defensible sites and fortifications more or less obsolete. During the Postpalatial period, when the Minoan civilisation was coming to an end, typical Minoan style defensive architecture was almost no longer constructed. Through time, the construction of guard houses and the methods used to modify access systems had lost their purpose (Alusik 2007, 148). In LM IIIB-C many new fortifications were constructed, consisting of massive walls and new guard houses (Alusik 2007, 154). However, from that period we start to enter the Cretan Dark Age (see Nowicki 2000, 41-222, for a catalogue of Dark Age sites), which is not the focus of this thesis.

As a final note on the significance of fortifications in general, Zielinski (1998, 528) stated that the construction of cyclopean fortifications of different kinds has played an important part in Minoan state formation. In a way it has indeed. The unrest during the Protopalatial period was followed by a defensive network of guard houses and watchtowers that helped create stable territories and states (for a while at least). Therefore, defensive structures seem to have had an important effect on how states were formed.

(20)

18 So far this thesis has provided a brief overview of the developments within Minoan archaeology relating to conflict and crisis on Minoan Crete. First the shift from a belief in a fully peaceful society to the acceptance that conflict and crisis were actually present. Then the discussion moved to the most important events that caused trouble within Minoan society and some possible effects, after which we took a closer look at the more prominent and surviving remains that provided evidence for unrest and conflict, namely fortifications. Having provided the necessary context, it is now possible to focus on the subject of this thesis: the Minoan guard houses.

The guard houses of Minoan Crete are an interesting type of architecture to study. Although they are usually quite small in size, there seems to be more going on than meets the eye. Arthur Evans saw the strong foundations of several guard houses and called them by different names, depending on their size and location. Examples are frouria (forts), guard stations and mother forts (Chryssoulaki 1999, 76). The name guard station or guard house stayed in use ever since, as is the implied function.

There are several features that have often been associated with guard houses. Although several specific guard houses will be discussed in detail in the next chapter, it is important to already give a clear idea of what is actually meant by a guard house. Chryssoulaki (1999, 78-81), who is part of the Minoan Roads research project, has presented a list of these features. This will be summarised below. It is important to keep in mind that not all features are present at every guard house.

1. Guard houses usually have a square plan of about 10x10 to 12x12m. If the building is smaller, it is either not completely preserved or it is a vigla, an outpost. The outer walls are constructed using megalithic masonry: large blocks of stone with small stones filling the gaps. These walls can be 1m to 1.20m thick. 2. Guard houses were always constructed with blocks of sideropetra, a grey

limestone that was locally available. The style of construction is very typical of the MM II period.

3. There seems to be a default way of how the interior of these structures was arranged, based on some excavated examples. The main room is a paved rectangular space, surrounded by the other rooms. This is not always the case however.

4. A terrace, in many cases at the point which offers the widest view and control over the surrounding area.

5. A side entrance with a low doorway is visible in some guard houses. This is called a sally port, because of the similarity with sally ports in Mycenaean context, but it is not certain at all.

(21)

19 6. It is thought that the roofs have been flat, but we cannot know for sure, since

roofs have not been preserved. The height is also a little problematic. The strong walls together with evidence for an upper floor suggest that they may have been buildings with multiple floors, although we cannot make any statements about their presumed height.

7. There are cases where there has been put a lot of effort in making it possible to build a guard house at a specific location. In those cases, foundations were constructed. Foundations were often very irregular, filling up ditches, levelling slopes, etc., to make it possible for a structure to be built on top.

8. Most guard houses were isolated structures in the landscape. It was not necessary for guard houses to be near any other site or settlement. The builders did their best to give the guard houses a very wide view, while making sure the guard house itself would be hard to access by hostiles.

9. Supplementary structures, to control access to the guard houses were built. Examples are strong enclosure walls that were sometimes placed in multiple rows around the guard house, and vigla, small look-out structures, that were placed near the guard house (although distance may vary). Sometimes these walls could be very high, as is the case at Mavro Avlaki, where the wall is preserved to a height of 2 m.

10. Guard houses were often built at natural features on low elevation that both give a good view of the surrounding area and are easily defended. Chryssoulaki sees the protection of roads as their main function. Cliff edges were often used, and when combined with walls, vigla and other natural features, these made guard houses very defensible.

11. Guard houses can be found in every part of the area studied by the Minoan Roads project. However, they do seem to be spread around main roads, but that was not yet confirmed when Chryssoulaki published these features.

12. In some guard houses sherds from the MM II period have been found, but LM I and LM III sherds have been found in all guard houses. Guard houses were often reused in later periods for all kinds of purposes. This later occupation can make it difficult to interpret some structures, since the sites have often been disturbed too much.

13. Some guard houses may actually not have been guard houses at all, but more like villas in the countryside. Their size, location and building quality make them stand out from the other guard houses.

The above points are the main features of Minoan guard houses. These can be further summarised into three general characteristics, namely 1) a clear defensive character, 2) a

(22)

20 shared architecture, and 3) a close relation to the road network, with a controlling view of the surrounding landscape (Chryssoulaki 1999, 81).

Guard houses are usually subdivided into three groups. Firstly, the central posts or administrative posts, the complex guard houses. These are the largest guard houses, often built from higher quality stone. They may also have functioned partially as caravanserai. Secondly, standard guard houses. This is most common type. These guard houses are very small and simple, possibly comprising a few small rooms. Thirdly, smaller watchtowers or vigla. These are very small towers that seem to have served as lookouts and possibly for communication over longer distances as well. Vigla are often connected to a complex or standard guard house by a peribolos (Alusik 2007, 126). Alusik also lists as guard houses other ‘cyclopean’ style buildings where the defensive function was probably secondary to its primary role, for example a strongly built farmstead. According to Alusik they qualify as guard houses if they are on the right place in the landscape (Alusik 2007, 124). However, in many cases the true function of this kind of structure is hard to prove, if at all possible.

As may have become clear already, the main function of guard houses was to control movement and watch over the surrounding area. It is seen as no coincidence then, that they were mostly constructed near important roads or stone quarries. Examples of guard houses near a stone quarry are Aspres Plakes, Chochlakies, Karoumes, Polla Kladia, Mavromouri and Chiromandres (Alusik 2007, 127).

Because most guard houses share the same architectural style, it is thought that many may have been part of a large building programme, in which also the main roads were constructed, at the end of the EM period or the beginning of the MM period (Chryssoulaki 1999, 81-2). Although most guard house sites have been occupied from at least MM II, some of the earliest guard houses have been dated to MM I, for example the guard house at Myrtos Pyrgos (Alusik 2007, 150).

During the Protopalatial period, more and more guard houses and similar structures were built. Their occurrence has been related to the emergence of larger palatial territories. These centres needed to define and protect their own hinterland, hence roads and defensive fortifications were constructed. The main function of this network of roads and guard houses was, as already mentioned above, to control access to roads, which was basically a military function (Alusik 2007, 129). The most important and largest guard houses have been dated to the Protopalatial period. Not only were these located at defensible locations, they also had perimeter walls around them. The latter not only offered more defence, but also were a statement of power (Zielinski 1998, 518).

During the Neopalatial period the role of guard houses becomes more complex. Many old guard houses were still in use and new guard houses were built, often on

(23)

21 locations that were not as defensible as before and without external fortifications, such as

periboloi (Alusik 2007, 129). These smaller guard houses were constructed all over the

territories to maximize controllability (Zielinski 1998, 519). However, at the same time, other guard houses were abandoned or adapted to fulfil a new function as farm, villa or workshop, e.g. at Sfaka. The above is probably related to a stable period without much unrest. Some scholars claim that Knossos united the island under one banner, which is still up for discussion, however, political stability, one way or another, would create a stable and safe society. Maybe because of this, guard houses were no longer needed as much to control access, but were instead used to watch over the economic activities, such as farming and crafting. These were necessary to support the palatial centres (Alusik 2007, 129).

Later, during LM II, not many new guard houses were built and the few that were built lacked the architectural qualities of the earlier periods. Also, vigla, periboloi and roads were lacking. It seems that at that moment the landscape was mostly no longer systematically controlled. The new guard houses that were constructed merely functioned as watch points or to control specific roads by themselves (Alusik 2007, 129), without being part of a defensive system.

Throughout the Cretan Bronze Age, guard houses seem to have functioned as control points, territorial defence and symbols of the status and power of the elite that was in control of the territory. All these functions can explain why these structures were mostly built on strategic locations (Zielinski 1998, 522-4). However, how defensible were these guard houses? Alusik (2007, 134-5) already noted, as a response to Zielinski, that although they were constructed using ‘cyclopean’ masonry, they would not last long against a real military force. Furthermore, he does not agree with the idea that during the Neopalatial period the Minoans changed to a system of border defence instead of town defence. Although there is evidence to support the idea of a territorial defence, which has been discussed already earlier in this chapter, Alusik has a good argument stating that this system would only work if the spread of guard houses across the island was regular, which is not the case. On the contrary, only in a few areas is there a high concentration of guard houses, and even in those areas the spread is not really regular (Alusik 2007, 134-5). There is a problem with Alusik’s argument however. The lack of evidence for guard houses in other areas of Crete is most likely the result of no available documentation for other parts of the island that have not been studied thoroughly yet. Compared to other areas of Crete, the eastern part of the island, especially the territory of Kato Zakros, has a very high concentration of defensible sites. Many of them have been discovered by the research team of the Minoan Roads Project. These do not only include guard houses, but fortification walls and strongly built farmsteads as well.

(24)

22 When reading through the literature, one finds a lot of references to the defensibility of the Minoan guard houses. However, it occurred to me that this was mostly based on 1) the thick walls, and 2) their location. Thick ‘cyclopean’ walls have almost always been interpreted as defensible. When looking at the variety of functions of the different types of buildings that were constructed using this type of wall, there are more possibilities. Walls can be made strong for a variety of reasons. Thick walls are not only a sign of fortifications, but can be a way to isolate buildings to counteract high temperatures during the summer and low temperatures during the winter.

Another aspect of the guard houses that should be explored is their accessibility. Many guard houses are situated at locations that are difficult to reach. This was either because of the location itself or because of additional walls surrounding the site. There does not seem to be any previous research discussing this thoroughly, but it deserved more attention, because the identification and interpretation of guard houses are largely based on their degree of accessibility.

Because of the above questions related to the location and defensibility of guard houses, the following main research question has been studied: How defensible were the Minoan guard houses? The results thereof will be presented in this thesis.

In this study, multiple research methods have been used to (try to) get a better idea of the defensibility of a select number of guard houses. See the next chapter for an overview of the case study. The first method that was used was Least Cost Path (LCP) analysis, using GIS and a 3-D landscape model. It would be logical to assume that a defensible building is hard to reach. You would not expect such a building to be approachable from all directions, which would make it easy to attack it. Rather the opposite should be the case. To determine how many routes one could take to reach a guard house, LCP analysis was performed on a DEM of the research area. For a detailed description of the method and how it was performed, see chapter 3. For the results and a discussion, see chapter 4.

The second method that was used, is a quantitative approach which results in a Defensibility Index. This is an adaptation of a method used by Canadian researchers that studied the defensibility of sites on the west coast of Canada. Although it may not give a straight yes or no answer to the question if a certain guard house is really defensible or not, it gives an indication of its defensibility compared to the other sites in the sample. More information on this method can be read in chapter 3, and the results are discussed in chapter 4.

The third and last method that has been used is Space Syntax. The above two research methods both study the exterior of these buildings. Space Syntax was used to study the inside of the selected guard houses. Whether a guard house is defensible from

(25)

23 the outside or not, the interior of a building can have an effect on its defensibility as well. In case hostiles get a chance to reach the guard house, there may still be a chance to defend it from the inside, or at least keep people inside more or less safe. Using Space Syntax theory and its Visibility Graph Analysis it was tried to get a first idea of the defensibility of the inside of guard houses. More information about this method can be found in chapter 3, with the results and a discussion in chapter 4.

Even though the results are already interesting, they must be looked at critically. Nonetheless, there may be room for improvement, but this thesis presents multiple new ways to study guard houses and it shows the possibilities for the future that may shine a new light on a topic that seems to have lost much interest in the last decade.

(26)

24 Figure 2: The structure plans of selected guard house that were studied (after Chryssoulaki 1999, plate VIIIa).

3: Plakalona 12: Chiromandres 4: Farangouli 14: Mavro Avlaki 5: Mavromouri 15: Polla Kladia 7: Chochlakies 16: Kali Elia 9: Kokkino Froudi 17: Aspres Plakes

10: Sfaka 18: Karoumes – Fort of the Sea 11: Agio Pnevma 19: Karoumes – Mother Fort

(27)

25

2 Case study

The previous chapter provided an overview of previous research and explained the goal of this thesis. Before continuing with the discussion of the methods of analysis that were used and the results, it is important to discuss the sample that was used for the study of the defensibility of Minoan guard houses. Figure 2 shows the selection that was made, except for three sites of which no building plan was available. Figures 3a and 3b (next page) show the area of Crete that has been studied and the locations of these guard houses on the map. This sample was mostly based on the publication of several guard house plans by Chryssoulaki (1999, plate VIIIa). A large number of these plans, however, did now have any clear entrance. It is possible that people entered from an upper floor, which has not been preserved for any of them. Also, many of the guard houses plans were too simple and/or too small to really say anything about the defensibility. At least, that was the idea before this study began. The guard houses that have been used in this research, are the ones with a more complex layout of those with an entrance, but several simpler buildings have been included as well. For the Least Cost Path analysis, the lack of an entrance, the low complexity of most of these guard houses and the lack of a building plan did not matter, so all were used. For determining a Defensibility Index, three guard houses were unsuitable for analysis, because there was no plan available to determine the size of the guard house site itself. The same smaller selection was used for the Space Syntax analysis, since this also requires good building plans.

Before providing more details about the selected guard houses, some information will be given about their discovery. Many Minoan guard houses have been discovered in the eastern part of Crete, in the area between Palaikastro in the north and Ambelos in the south. Although there have been multiple research projects in this area of Crete, the most noteworthy may be the Minoan Roads project (see Tzedakis et al. 1990; Tzedakis et al. 1989). This is a project of the Greek Ministry of Culture that was started already in 1984, with the goal to study the layout and construction of roads and a road network during the Protopalatial period. The project first started in western Crete, but moved to eastern Crete in 1986 when it became possible to date roads (Tzedakis et al. 1989, 45). Based on the assumption that constructing roads takes a lot of effort and required a high level of organisation and a large number of people, the project also included a study of Minoan society in this period (Chryssoulaki 1999, 75).

(28)

26 Figure 3a: Map of Crete with the research area marked.

(29)

27 The eastern part of Crete was chosen as the main research area, and in the first half of the 1990s this area was systematically surveyed. In addition to the study of the road network, it was also decided to map any archaeological sites that were discovered. Hence, the Minoan Roads projects did not only contribute to our better understanding of Minoan roads, their function, dating, and conservation (Chryssoulaki 1999, 75), but also of other sites, among which were many guard houses, vigla and related structures. Much of this knowledge has already been discussed in the previous chapter.

Tzedakis et al. (1989, 44) recognised three development stages in the construction of a road network. The first roads are very simple roads meant for communication. The second type of roads is meant for vehicles. These roads are of better construction, are wider, less steep and can be paved. The third type of roads is military roads. The latter are the most important roads that can cross a large region and extent over long distances. These roads are usually straighter, use easier slopes and are guarded by guard houses. This military function of the road network was also attested by Tzedakis et al. (1989, 60). Moreover, this has been used by some to explain the presence of Minoan guard houses. Zielinski, for example, suggests that the road system was constructed at the same time as the network of guard houses, beginning in the Protopalatial period with a peak during the Neopalatial period, as a means to claim territories (Zielinski 1998, 538). It is during this period of development of society that guard houses held multiple important functions. Not only did they control access to roads, they also monitored the entire region by being visually interconnected (Alusik 2007, 131-2). Based on the correlation between many guard houses and the road network, it can be established that guard houses seem to be placed in relation to specific roads, even if they were not constructed at the same time. Many guard houses that were found and surveyed during the project have been published by the Minoan Roads research team, although not all publications are easily available.

Now that the history of the project responsible for the documentation of most guard houses in eastern Crete has been discussed briefly, each of the guard houses in the sample will be introduced individually. It must be noted here, that not all of these guard houses have been described in detail in publications (or publications were unavailable). In many cases, there is little information available. However, especially for those guard houses, the results of this defensibility research are interesting. The sites will be discussed in alphabetical order. When no size information was available, the buildings were measured in AutoCAD.

Agio Pnevma: this site is located on top of a flat hill that looks over the surrounding area (Alusik 2007, 126). Similar to other guard house sites, there are one or more periboloi around the site following the edges of the hill (Alusik 2007, 131). This building consists of only one large room of about 11x15 m in size.

(30)

28 Aspres Plakes: this site is a good example of a guard house close to an important source of stone, possibly to control the quarry (Alusik 2007, 127; Zielinski 1998, 431). This guard house, which measures roughly 8x10.5 m, consists of three spaces, although it is difficult to establish whether these functioned as individual rooms. The inside is subdivided by two small walls.

Chiromandres: the site of Chiromandres has been described as a guard house, based on the topography and the archaeological finds at the site. The structure, which is about 14x28 m in total, has two building phases, one dating to MM II and one dating to the Neopalatial period. Zielinski thinks, based on comparable sites, that this structure was multifunctional, which would set it apart from other buildings that only have the guard house role (Zielinski 1998, 505). In the Neopalatial period, many guard houses were adapted to fit other purposes, so that could explain the multifunctionality that seems to be visible at Chiromandres. Tzedakis et al. (1989, 55) suggest the road which leads from Kato Zakro south to Ambelos passes close to this site. The middle of the building has a paved floor and there are several rooms around it. Evidence suggests a possible role as a kind of service station, because of the available kitchen, water supply and discovered millstone. These remains have been dated to the Neopalatial period and later (Tzedakis et

al. 1990, 51).

Pottery finds at the site suggest two periods of occupation, MM IIA and MM IIIB/LM IA. This fits the two construction phases and is also similar to the dating of other guard houses and is possibly connected to the start of the Minoan road network (Tzedakis

et al. 1989, 72; 4). Sometime during LM I this guard house seems to have been

abandoned, but it was reoccupied again during LM III, possibly only for a short time (Alusik 2007, 153; 5).

The guard house is located at the north-eastern side of a plateau, where it has a good view overlooking the valley below. Around it are walls and watchtowers, but these are much smaller than the guard house itself (Tzedakis et al. 1990, 48). From the

periboloi around the guard house, only the foundations have been preserved (Tzedakis et al. 1990, 56). At Chiromandres, the walls that surround the terraces around the site were

large and connected to the main building. Alusik suggests they may have functioned as ramparts (Alusik 2007, 131). The nine vigla were also connected to the guard house by

periboloi (Alusik 2007, 140; 50).

Chochlakies: this is another example of a guard house that was located close to an important stone quarry in the south-western part of the plain of Chochlakies. It was one of the later guard houses, being constructed in LM I (Alusik 2007, 127-8). This building is about 10x10 m in size and is located on a terrace that is difficult to access. According to Zielinski (1998, 504), this guard house takes a bastion position in the landscape.

(31)

29 Farangouli: there was no information found about this guard house in the available literature. However, it was an L-shaped guard house, approximately 10x12.5 m in size.

Kali Elia: this guard house is also located on a defensible terrace, situated above the surrounding area, with a good view towards the west from a bastion-like position (Alusik 2007, 126; Zielinski 1998, 504). There were periboloi following the edges of the hill around the building (Alusik 2007, 131).

Karoumes – Fort of the Sea: this is another example of a guard house close to a quarry. It is located at the eastern end of the Chochlakies gorge, with a good view of the coastal plain and the pass through the gorge. Because of its large size of 12x20m, this guard house was interpreted as an important administrative centre (Zielinski 1998, 502). It was built during MM IIA and was already abandoned during MM III (Alusik 2007, 127-8). Several vigla and periboloi were constructed near the site at the same time (Alusik 2007, 150). During LM I the guard house was occupied again and a new construction phase is visible. However, in LM IB, it was destroyed by fire (Alusik 2007, 128). Alusik suggests there was no entrance found at this guard house (Alusik 2007, 130), whereas the published plan of the building shows an entrance. There are indications that this site served multiple functions. Besides the function as a guard house, there is evidence for farming activity as well (Zielinski 1998, 432).

Karoumes – Mother Fort: this building is located very close to Karoumes – Fort of the Sea. Just like the Fort of the Sea, this guard house is located near a quarry (Alusik 2007, 127). Karoumes – Mother Fort was also constructed during MM IIA. Unlike the Fort of the Sea, this guard house remained in use during MM III. Here, too, periboloi and

vigla were located around the site (Alusik 2007, 150; 2). The walls were large and

connected to the guard house. Terraces had been constructed in front of the guard house, possibly as watch points, but maybe also to build a workshop or stable on (Alusik 2007, 130-1). It was approximately 10x13 m large.

Kokkino Froudi: this guard house was constructed during MM IB/II on a high point on the edge of a gorge. The guard house was surrounded by walls which possibly functioned as ramparts. During LM IA this guard house was adapted to serve as a pottery workshop (Alusik 2007, 126; 8; 31; 50). The building was approximately 9x13 m large.

Lidoriako: there is not much information found about this site in the literature. It is a vigla along a road towards Ambelos (Tzedakis et al. 1989, 75), located on a high point (Alusik 2007, 126). There is no plan of this building, hence the dimensions are unknown.

(32)

30 Mavro Avlaki: this is another guard house of which not much is published. Apparently, there is still a 2m high wall preserved around this guard house (Chryssoulaki 1999, 79-80).

Mavromouri: the only available information states that this building was also located very close to a stone quarry (Alusik 2007, 79-80). As can be seen from the plan, it seems to have multiple rooms that are not connected to each other and do not seem to have any entrances. Again, it may have been possible to enter from an upper floor. This guard house was about 10x12.5 m large.

Melefa: there was no plan of this guard house available, but according to Tzedakis et al. (1989, 75) this guard house was located along the transportation route to Ambelos.

Plakalona: other than the plan of the guard house, there was no information available. This structure is about 10x12.5 m large and has one room, which is subdivided by a short wall.

Polla Kladia: this is another guard house that is located on a terrace on a hill, close to a stone quarry, looking over the surrounding landscape (Alusik 2007, 126-7). As can be seen from the plan, there is one large room and one small room, not connected to each other and without any entrance. Again, it may have been possible to enter from above. The guard house was approximately 12x16 m large.

Pyrgales: this guard house was constructed during MM IB and it seems to have remained in use until it was destroyed in LM I (Alusik 2007, 128; 50-52). Although no building plan was available, it has been suggested that this was an important administrative post for the upland areas (Tzedakis et al. 1989, 75; Zielinski 1998, 452), which might suggest a more substantial layout and larger dimensions.

Sfaka: this guard house was constructed in MM IB at a high point on the edge of a gorge, but it was already abandoned at the end of MM II. During the Neopalatial period, this guard house was reoccupied and adapted to function as a workshop (Alusik 2007, 126; 9; 50-51). Zielinski only suggests a Neopalatial date, but does note the existence of a

vigla at the site as well, and that this site was located on a ridge along the eastern side of

the road from Zakros to Ambelos, possibly guarding the route (Zielinski 1998, 504-5). The above listed sites are the guard houses that have been studied. For most of them a building plan was available, allowing all of the analyses to be carried out, while for the three guard houses without a plan only the Least Cost Path analysis was conducted. The following chapter will discuss the different analyses that have been performed on these guard houses, followed by a chapter presenting the interpretation of the results and the interesting insight that can be inferred.

(33)

31

3 Methodology and methods

In the last chapter, the sample of guard houses was discussed. In this chapter, the research methods are discussed and it will be explained how they have been applied to the selected guard houses from the Zakros area in the far east of Crete. This chapter is divided as follows. First the methodology foreach type of analysis will be discussed. This is followed by a description of the specific methods and tools used for the analyses. Before continuing, however, it is important to restate the goal of the research.

The goal of the research was to examine the defensibility of Minoan guard houses, by conducting Least Cost Path analysis, calculating a Defensibility Index, and carrying out Visibility Graph Analysis. By applying methods that have never been tried out before on these buildings, it may be possible to not only give new insights on the guard houses themselves, but also critically evaluate the methods that were used and to explore how to further improve them. The former may rekindle an interest in these structures and any other type of architecture that has been called defensive or defensible. The latter will show examples of methods that can be perfected in the future to widen our understanding of guard houses and similar structures.

3.1 Methodology

3.1.1 Least Cost Path analysis

In this section, the methodology behind the methods will be discussed in the order they have been applied, starting with Least Cost Path (LCP) analysis. To better understand the use of LCP analysis for studying guard houses, it is important to first discuss the relevant GIS background.

GIS software packages developed in the last one and half decades or so have made it a lot easier for people to analyse landscapes. While these packages were first used mostly by geographers, archaeology has been using it more and more for a long time now as well. GIS allows for the reconstruction of landscapes based on numerous input data. Most often Digital Elevation Models (DEMs) are used to construct the shape of the landscape, while other datasets may add vegetation, land use, etc. Another advantage that is very useful for archaeologists, is the ability to add a time factor if needed. The possibilities are almost endless.

For this study, Least Cost Paths were calculated in ArcGIS. The exact method that was used will be discussed in the Methods section, but first it is important to briefly explain what Least Cost Path (LCP) analysis is and what the (dis)advantages of this method are.

Defensible or not defensible? Guard houses from Middle and Late Minoan Crete revisited, using GIS and Space Syntax

Defensible or not defensible?

Guard houses from Middle and Late Minoan Crete

revisited, using GIS and Space Syntax

Defensible or not defensible?

Guard houses from Middle and Late Minoan Crete revisited, using GIS and

Space Syntax

Table of contents

Acknowledgements

1 Introduction

2 Case study

3 Methodology and methods

3.1 Methodology